Research On Generalized Predictive Control For Wind Tunnel Mach Number

With the development of large aircraft,the control accuracy of Mach number control of wind tunnel test which provides important basis for the development of aircraft is put forward,but the existing control methods in long regulation time,overshoot is the Mach number in the test,the problem of high cost of blowing tests.Considering the future predictive optimal control based on moment control deviation rate,has the ability to achieve high accuracy and fast control.Therefore,this thesis takes an intermittent transonic wind tunnel as the plant and researchs using predictive control to realize fast and high precision of wind tunnel flow field control.By means of analysis,the generalized predictive control algorithm with a wider range of application is chosen to realize the control of the high accuracy of the wind tunnel flow field.The key to the realization of this algorithm lies in the establishment of the prediction model,which is the determination of the model structure,the order of the time delay and the adjustment of the relevant parameters in the algorithm:(1)A good model structure can describe the actual characteristics of the object,to lay the foundation for the realization of the fast high precision control.In the wind tunnel flow control structure,the main pressure loop is introduced,which can reduce the control difficulty and eliminate the time-varying effect caused by the pressure drop.The controlled quantity is selected as total pressure and static pressure,which reduces the system nonlinearity of direct control of Mach number.Because the flow field near the operating point can be approximated as a linear system,the linear model is chosen.Compared with the CARMA model,the CARIMA model has the effect of eliminating the static difference and is closer to the actual object characteristic.The ability of approximate decoupling can reduce the influence of wind tunnel coupling,so choose two diagonal CARIMA models as the structure of wind tunnel flow field prediction model;(2)The appropriate order is conducive to data mining,so that the prediction results are more accurate.In this study,the model order is determined by using the false nearest neighbor method.The percentage of the false neighborhood is given by the method,and the model order is obtained when the percentage is close to 0,which is simple and easy to implement;(3)On the one hand,it is beneficial to improve the precision of model parameter identification,on the other hand,it provides a reference for the whole set of minimum prediction time domain N1.Method of using first order plus dead time model with least squares curve fitting to determine the time delay,the method can rapidly determine the object of delay analysis characteristics of wind tunnel flow field with time delay and quantitative;(4)the parameter tuning is directly related to the performance of the control system,but there is no general rule,according to the parameters of first order plus dead time model tuning method based on previously proposed and actual debugging experience can realize parameters,this method greatly simplifies the difficulty of tuning parameters and it has achieved satisfactory control effects and in practical application.In the wind tunnel control system simulation platform based on Lab VIEW builds on point near Mach number test,compared with PID control,the controller based on generalized predictive control algorithm under the condition of meet the control precision of both reduce the overshoot and shorten the time is up to the expectations of adjustment.Due to the dynamic characteristics of the flow field in wind tunnel with operating mode is changed,the variable Mach number experiment was carried out between the different working conditions,the model change is bound to impact on high precision control,in order to eliminate the influence makes the controller to adapt to this change,the controller is improved,the introduction of with forgetting factor recursive least square method to overcome the traditional recursive least squares method new data on parameter estimation problem of weak influence of realize the adaptive control,the flow field in wind tunnel simulation results verify the effectiveness.